Five healthy male subjects and one healthy female subject (mean [ ± SD] age, 33 ± 6 years) participated in the study. None was taking any medication or had any remarkable medical history. All had at least average fitness; the mean (± SD) maximal oxygen uptake was 50.0 ± 12 ml Vkg/min. further

**Exercise Testing**

Maximal exercise tests were performed using a ramp treadmill protocol, ie> using a constant, continuous change in external work. Initially, an exercise test was performed to familiarize subjects with the procedure and to determine maximal oxygen uptake. Using this information, treadmill ramp rates were individualized to attain a test duration of approximately ten minutes. The same ramp rate, time of day, and laboratory conditions were employed for a given subject on day 2, and tests were performed a mean of 15 ±8 days apart. Ramp rates ranged between 1.2 and 7.0 ml Oj/kg/min, with a mean of 4.0 ±2 ml (Vkg/min, expressed as change in external work. The software and hardware used to perform these tests were developed by the Burdick Corporation (Milton, Wis). Exercise was continued to volitional fatigue and subjects were encouraged to give a maximal effort. The Borg 6 to 20 scale was used to quantify effort.

**Gas Exchange**

Pulmonary gas exchange variables were determined continuously throughout the exercise tests using the Medical Graphics Corporation 2001 System. Quality control documentation was performed using techniques outlined by Jones and Campbell. Data were sampled using 30 consecutive eight-breath average recursive sums. A single eight-breath average sample represented an average of the current breath and the seven preceding it. Thus, sampling was analogous to the average of 30 breaths, each breath representing an average of eight. The estimations of slope were made from each of these samples. The use of these averaging techniques has the effect of filtering variation. Figure 1 illustrates the averaging techniques used from which the estimations of slope were generated.

**Statistics**

The Students t test for paired observations was used to assess differences in perceptual, heart rate, and gas exchange responses between day 1 and day 2. To evaluate the slope of the regression lines for each individual ramp test, a BASIC computer program was developed to calculate the following: the slope of the change in oxygen uptake over time, the confidence interval and standard deviation of the slope, the calculated and critical T statistic, the Y intercept, and regression and residual sum of the squares.

*Figure 1. Gas exchange sampling technique used to calculate oxygen uptake slope.*